Commission rates don't move sellers. Margin math does.

The offer existed. The earning potential didn't.

Campaign messaging reached sellers by email only. No product surface existed — no way to check which listings qualified, no margin context, no self-serve data. Adoption was slow, but without any in-product signal it was impossible to know whether the problem was the offer, the messaging, or the missing interface.

Before building a UI, test whether sellers will engage with margin data at all.

There was pressure to build the full in-product experience immediately. My argument: we didn't know if the problem was the missing UI or the offer itself. The CSV was a cheap validation: qualifying listings and margin calculations in a downloadable format, added as a secondary action on a campaign banner. If nobody downloaded it, we'd know the interface wasn't the gap.

The Deals banner with CSV download. Deliberately low-friction: visible to sellers who wanted the data, but not promoted.

Campaign participation shouldn't feel like a one-time opt-in. It should feel like momentum.

The CSV confirmed sellers would engage with financial data. The next step was bringing that data into the product, but not as a static table. The research showed sellers needed two things: visibility into what deals were available and where they stood, and a sense of progress toward better commission rates. The deals dashboard introduced a dedicated surface inside the seller back office, a drawer accessible from the listings page that showed each active campaign as a card with real-time progress tracking.

The gamification was intentional but restrained: progress bars and milestone thresholds, not badges or leaderboards. The goal was to make the earning trajectory visible, not to manufacture urgency. Sellers could see exactly how many units separated them from the next commission reduction, and make their own call about whether to push for it.

From "you can earn more" to "here's exactly what this listing is worth under campaign terms."

The deals dashboard made campaigns visible and trackable. But "your commission drops from 9% to 8%" still didn't answer the question sellers were actually asking: is it worth adjusting my pricing for this specific listing? Each listing has a different sourcing cost, return rate, and volume expectation. A percentage improvement means something different for every item in a catalogue.

The tool calculates whether campaign participation yields better net earnings than a seller's current approach, per listing, based on actual pricing history and volume data. Two design decisions shaped how it presents that calculation. First, a table rather than individual listing views: sellers manage catalogues, not single items, and the format needed to support scanning across dozens of rows at once. Second, a three-state recommendation: Participate, Review, or nothing. "Don't participate" was deliberately excluded. The research finding from the Problem phase held here: sellers' resistance to coercion meant the interface could surface a recommendation but couldn't make the decision for them. "Review" flags the edge cases without closing them off.